Table 4.

Multivariable-adjusted odds ratio and 95 % confidence interval for association of levels of psychological problems profiles scores across the categories of breakfast consuming frequency^a

Psychological problems profile scores tertiles	Categories of breakfast consuming frequency			P trend
	Seldom (n 205)	Sometimes (n 424)	Always (n 2247)
Crude model				<0⋅0001
Lowest tertile	1 (reference)	1 (reference)	1 (reference)
Intermediate tertile	1⋅42 (0⋅96, 2⋅11)	1⋅18 (0⋅91, 1⋅55)	1 (reference)
Highest tertile	3⋅59 (2⋅52, 5⋅11)	2⋅20 (1⋅72, 2⋅82)	1 (reference)
Model 1				<0⋅0001
Lowest tertile	1 (reference)	1 (reference)	1 (reference)
Intermediate tertile	1⋅57 (1⋅01, 2⋅46)	1⋅23 (0⋅92, 1⋅64)	1 (reference)
Highest tertile	3⋅79 (2⋅52, 5⋅71)	2⋅33 (1⋅77, 3⋅08)	1 (reference)
Model 2				<0⋅0001
Lowest tertile	1 (reference)	1 (reference)	1 (reference)
Intermediate tertile	1⋅63 (1⋅02, 2⋅62)	1⋅22 (0⋅90, 1⋅65)	1 (reference)
Highest tertile	4⋅07 (2⋅64, 6⋅26)	2⋅38 (1⋅78, 3⋅17)	1 (reference)
Model 3				<0⋅0001
Lowest tertile	1 (reference)	1 (reference)	1 (reference)
Intermediate tertile	1⋅55 (0⋅96, 2⋅51)	1⋅24 (0⋅91, 1⋅68)	1 (reference)
Highest tertile	3⋅48 (2⋅20, 5⋅50)	2⋅37 (1⋅74, 3⋅22)	1 (reference)
Model 4				<0⋅0001
Lowest tertile	1 (reference)	1 (reference)	1 (reference)
Intermediate tertile	1⋅52 (0⋅94, 2⋅45)	1⋅18 (0⋅86, 1⋅61)	1 (reference)
Highest tertile	3⋅35 (2⋅11, 5⋅32)	2⋅25 (1⋅65, 3⋅06)	1 (reference)

^aMultinomial logistic regression model was used.

Model 1: Adjustment was made for age, sex, marital status and education. Model 2: Additional adjustment was made for body mass index, smoking and physical activity. Model 3: Psychotropic medicines use and gastrointestinal disorders were additionally adjusted. Model 4: Additional adjustment for dietary intakes including magnesium, riboflavin, pyridoxine, folate, cobalamin, Docosahexaenoic acid and Eicosapentaenoic acid, energy, fibre and caffeine.